Nitin Arora

203 Pages·2016·9.76 MB·English

by  Windows User

Mechanisms of Aggregation and Separation of Water and Solids from
Bitumen Froth using Cluster Size Distribution

by

Nitin Arora

A thesis submitted in partial fulfillment of the requirements for the degree of

Master of Science

in

CHEMICAL ENGINEERING

Department of CHEMICAL AND MATERIALS ENGINEERING
University of Alberta

© Nitin Arora, 2016
Abstract
The large-scale corrosion and equipment damage caused by water and solids
respectively in bitumen froth necessitate their removal using various methods. This study
is aimed at understanding how water and solids aggregate (water-water, water-solid and
solid-solid) in bitumen froth and what the dominant aggregation/ settling mechanism is
during gravity separation (coalescence, flocculation and sweep flocculation).
Initially, we compared two mixing/ settling tank designs: one with side sampling
ports and another with top sampling ports, to ensure that our sampling method is robust
and gives a representative sample. The side sampling design gave more meaningful
results for the two test emulsions. Bitumen froth contains a large amount of water and
solids. Hence, a robust image analysis algorithm was developed to quantify the type of
clusters (water-water or water-solid), the cluster size and the number of drops and/or
particles in the cluster. Using this clustering algorithm, it was found that water drops
flocculate with each other and also sweep flocculate the nearby solids. A low water
concentration in the product layer also ensured low solids concentration. Solid-solid
aggregates were rarely observed. The change in water drop size distribution over 60 min
of settling indicated some evidence of coalescence. Coalescence was also visually
observed for free water which could happen over hours or even days. Hence, flocculation
and sweep flocculation are both important settling mechanisms for the demulsifier used
in this study. Changing the mixing conditions (demulsifier injection concentration and
mixing energy) did not change the dominant settling mechanism. Good mixing promotes
aggregate growth, resulting in faster settling and lower final water and solids
concentrations, both of which are crucial for industrial operations.
ii
Acknowledgements
I would like to thank Dr. Suzanne Kresta for believing in me and guiding me
throughout the program. I want to thank Colin Saraka for providing support with
experiments, meaningful discussions and great team work. Many thanks to Akorede
Awosemo, Michaella Chemello for help in conducting experiments and Márcio Machado
for providing guidance from time to time. I want to thank Alena Kukukova for allowing
me to use her PNN Matlab code. Herb Green, from machine shop, deserves special
appreciation as he provided quick solutions and always find time for me.
I want to thank our project sponsors, Samson Ng and Sujit Bhattacharya, from
Syncrude Research who always found time to meet us in person, review my research
findings and provided great feedback. They also shared their research findings with us.
Finally, I would like to thank my family who has strengthened me throughout my
life. My father, Sushil Kumar Arora and my mother, Neeru Arora always encouraged me
to achieve great heights. My wife, Swapnali Shende, has been a strong guiding partner
and is indeed my better half.

iii
Table of Contents
Chapter 1: Mixing, Settling and Oil Sands Extraction ................................................. 1
1.1.1 Froth Treatment ................................................................................................. 2
1.1.2 Emulsion and Emulsion Stability....................................................................... 3
1.1.3 Demulsifier ........................................................................................................ 5
1.2 Mixing and Settling................................................................................................... 6
1.2.1 Mixing Characterization .................................................................................... 6
1.2.2 Meso-Mixing.................................................................................................... 10
1.2.3 Settling Mechanism ......................................................................................... 10
1.2.4 Sedimentation .................................................................................................. 15
1.2.5 Froth Mixing Studies ....................................................................................... 16
1.3 Research Objective ................................................................................................. 16
References ..................................................................................................................... 18
Chapter 2: Comparison of Sampling Orientation for Water/ Solids Settling
Experiments in a Diluted Bitumen System ................................................ 24
2.1 Abstract ................................................................................................................... 24
2.2 Introduction ............................................................................................................. 24
2.3 Experimental ........................................................................................................... 25
2.4 Results ..................................................................................................................... 28
2.4.1 Standard Deviation........................................................................................... 29
2.4.2 Full Profile Analysis of Water Concentration ................................................. 30
2.4.3 Microscope Results .......................................................................................... 33
2.5 Conclusion .............................................................................................................. 37
Nomenclature ................................................................................................................ 37
Acronyms ...................................................................................................................... 38
References ..................................................................................................................... 38
Chapter 3: Clustering Analysis for Characterization of Flocs/ Agglomerates in
Bitumen Froth Images ................................................................................. 40
3.1 Introduction ............................................................................................................. 40
iv
3.2 Image Analysis Methods......................................................................................... 45
3.2.1 Point Nearest Neighbor (PNN) ........................................................................ 50
3.2.1.1 Introduction ............................................................................................... 50
3.2.2.2 PNN Analysis Set Up ................................................................................ 52
3.2.2.3 PNN Results .............................................................................................. 55
3.3 Clustering Algorithm Steps..................................................................................... 61
3.3.1 Pre-Processing of Images ................................................................................. 61
3.3.2 Removing Fines and Separating Touching Objects ......................................... 64
3.3.3 Separating Water and Solids from Pre-Processed Image ................................ 69
3.3.4 Marking Centroids ........................................................................................... 71
3.3.5 Detecting Clusters ............................................................................................ 72
3.3.6 Finding Cluster Information ............................................................................ 74
3.3.7 Results .............................................................................................................. 75
3.3.8 Strengths and Limitations ................................................................................ 80
3.4 Conclusions ............................................................................................................. 80
References ..................................................................................................................... 81
Chapter 4: Settling Mechanisms in Bitumen Froth ..................................................... 85
4.1 Experimental ........................................................................................................... 85
4.1.1 Premixing ......................................................................................................... 87
4.1.2 Naphtha Blending (Froth + Naphtha) .............................................................. 88
4.1.3 Demulsifier Dispersion [(Froth + Naphtha) + Demulsifier] ............................ 89
4.1.4 Sedimentation .................................................................................................. 91
4.1.5 Sampling Schedule........................................................................................... 91
4.1.6 Experimental Design and Hypothesis .............................................................. 92
4.2 Results ..................................................................................................................... 94
4.2.1 No Demulsifier Run Results ............................................................................ 95
4.2.2 Factorial Design Run Results........................................................................... 96
4.2.2.1 Qualitative Image Analysis ..................................................................... 100
4.2.2.2 Clustering Image Analysis ...................................................................... 106
4.2.2.3 Drop Size Distribution and Coalescence ................................................ 117
4.2.3 Circulation Pattern Run Results ..................................................................... 122
v
4.2.4 OWS and CPA Analysis ................................................................................. 128
4.3 Conclusion ............................................................................................................ 133
4.4 Future Work .......................................................................................................... 135
References ................................................................................................................... 136
References (All) ............................................................................................................. 139
Appendix A: Experimental Data ................................................................................. 149
Appendix B1: Repeatability Experiments (Side vs. Top Sampling) for Diluted
Bitumen .................................................................................................. 155
Appendix B2: Repeatability Experiments (Side vs. Top sampling) for Water in
Mineral Oil ............................................................................................ 159
Appendix B3: Bitumen Froth Experimental Procedure ........................................... 162
Appendix C1: Standard Operating Procedure for Sending OWS/CPA/EXM Samples
to Syncrude ............................................................................................ 167
Appendix C2: Standard Operating Procedure for Receiving Feed Material from
Syncrude ................................................................................................ 172
Appendix C3: Lab floor Cleaning and General Housekeeping ................................ 174
Appendix D1: Karl Fischer Procedure ....................................................................... 176
Appendix D2: Microscope Image Acquisition Procedure ......................................... 182

vi
List of Tables
Table 1-1: Power numbers (N ) for various impellers in the fully turbulent regime with 4
p
standard baffles in a standard geometry stirred tank [adapted from
(Hemrajani and Tatterson, 2004)] .................................................................... 8

Table 2-1: Repeatability test results of water content for top and side sampling methods
at 3 sampling heights (z2, z3 and z4) and two liquid-liquid systems ............... 30
Table 2-2: Water content of samples (wt %) collected from 4 different heights during a
60 min settling period at poor mixing conditions in top sampling CIST for
diluted bitumen .............................................................................................. 36

Table 3-1: Repeatability of index of dispersion (I) for random data generated for object A
and object B using rand ( ) function .............................................................. 56
Table 3-2: Normalized index of dispersion, as a measure of clustering for several
clustering scenarios (Figure 3-8 and Figure 3-9) for object A and object B . 57
Table 3-3: Effect of cluster orientation on index of dispersion for object A (when B is
randomly dispersed) for different grid sizes .................................................. 59
Table 3-4: Effect of cluster spacing on normalized index of dispersion for object A
(when B is randomly dispersed) .................................................................... 59
Table 3-5: Effect of percentage of clustered objects on index of dispersion ................... 60
Table 3-6: Example of connectivity matrix: First row (1, 2) indicates object number 1
and 2 are connected and so on ....................................................................... 73

Table 4-1: Composition of supplied bitumen froth .......................................................... 87
Table 4-2: Premixing can geometry and mixing parameters ........................................... 88
Table 4-3: Mixing specifications for naphtha blending (Froth + Naphtha) step .............. 89
Table 4-4: Mixing conditions for demulsifier dispersion step ......................................... 90
Table 4-5: Operating conditions and run summary for bitumen froth experiments using
Rushton (RT), A310 and Intermig (IM) impellers ......................................... 91
vii
Table 4-6: Variable range for demulsifier dispersion in bitumen froth experiments using
Rushton (RT), A310 and Intermig (IM) impellers ......................................... 93
Table 4-7: Comparison of Rushton and A310 runs at same mixing conditions ............ 127

List of Figures
Figure 1-1: Illustration of water-in-oil emulsion (left) and oil-in-water emulsion (right)
with emulsifier molecules shown as jagged lines ............................................ 4
Figure 1-2: Schematic of a stirred tank with impeller diameter (D), impeller off-bottom
clearance (C), tank diameter (D), liquid height (H) and impeller rotational
speed (N) .......................................................................................................... 7
Figure 1-3: Three steps of coalescence: drop collision, film drainage and film breakage
....................................................................................................................... 11
Figure 1-4: Sweep flocculation happens when a faster settling floc surpasses small drops
which then become part of the floc ................................................................ 14

Figure 2-1: The geometry for side (left) and top (right) sampling orientations in the CIST
....................................................................................................................... 26
Figure 2-2: Experimental set up showing side (Left) and top (Right) sampling CIST
along with needles used for withdrawing samples ........................................ 28
Figure 2-3: Microscope slide preparation involves putting a drop of liquid on to silanized
microscope slide using a Pasteur pipette and covering it with a cover slip
(image source: creative commons)....28
Figure 2-4: Water content of samples collected from 4 different heights during a 60 min
settling period at favorable mixing conditions for water in diluted bitumen
with (a): Side sampling CIST (Refer data: run #RB, Table A-1) and (b): Top
sampling CIST (Refer data: run #RA, Table A-1) .......................................... 31
Figure 2-5: Water content of samples collected from 4 different heights during a 60 min
settling period using side and top sampling needles from same CIST for water
in mineral oil (Refer data: Table A-3) ........................................................... 32
viii
Figure 2-6: Diluted bitumen microscope image captured with 40x lens at favorable
mixing conditions, (a): no aggregates in feed can sample after premixing, (b):
solid-water aggregates 60s after demulsifier addition at z in top sampling
1
CIST, shown in rectangular box..................................................................... 33
Figure 2-7: Diluted bitumen microscope image captured with 10x lens at favorable
mixing conditions in side sampling CIST, (a): showing solid-water aggregate
for sample (5 min at z ), (b): aggregates were absent for sample (7 min at z )
2 2
....................................................................................................................... 34
Figure 2-8: Diluted bitumen microscope image showing solid-water aggregates at
favorable mixing conditions for sample (5 min at z ) in side sampling CIST,
2
(a): 10x lens with solid-water aggregate in rectangular box, (b): 40x lens at
same location as image (a), with arrows indicating water drops trapped in the
aggregate ........................................................................................................ 35
Figure 2-9: Diluted bitumen microscope image captured with 40x lens, showing loosely
packed solid-water aggregates at poor mixing conditions ............................. 36

Figure 3-1: Illustration of hollow object (left) such as water and filled object (right) such
as solids on a microscope image. Their shapes are frequently non-spherical.
....................................................................................................................... 43
Figure 3-2: Representative diluted froth images captured with a 40x lens containing (a):
water-water flocs and (b): solid-solid chains ................................................. 43
Figure 3-3: Representative diluted froth images captured with a 40x lens containing (a):
water-solid flocs with water drops indicated by an arrow and (b): water-water
chains ............................................................................................................. 43
Figure 3-4: Representative diluted froth images showing examples of (a): free water
covered with solids and bitumen clay skins, (b): free water containing trapped
bitumen drop (rectangle) and bitumen film (arrow) at interface, (c): free
water aggregates ............................................................................................ 45
Figure 3-5: Diluted froth image showing (a): water-water floc and (b): water-water
chains. Nearest neighbor distance of Figure (a) and Figure (b) are shown in
ix
Figure (c) and Figure (d).Nearest neighbor direction (radians) of Figure (a)
and Figure (b) are shown in Figure (e) and Figure (f). .................................. 47
Figure 3-6: (a): Dendrogram for image shown in Figure 3-5a with object set number and
their separation distance (µm) on horizontal and vertical axis respectively,
(b): Two group of clusters identified with dendrogram, one with yellow
circles and other with green arrows ............................................................... 49
Figure 3-7: Point Nearest Neighbor method showing single hexagon in a grid. Each grid
point (black circles) looks for nearest water drop (red circles) and solid (blue
rectangles) in image space [Modified significantly from Kukukova et al.
(2011)] ........................................................................................................... 51
Figure 3-8: Test images containing two objects: object A (red) and object B (green) with
hexagonal grid points (blue) shown for three scenarios, (a): A and B
randomly distributed, (c): A aggregated but B randomly dispersed and (e): A
aggregated and B aggregated separately. The point-nearest object distribution
for Figures (a), (c) and (e) is shown in Figures (b), (d) and (f) respectively. 53
Figure 3-9: Test images containing two objects: object A (red) and object B (green) with
hexagonal grid points (blue) shown for three scenarios, (a): A and B
aggregated together, (c): A aggregated in long vertical chains but B randomly
dispersed and (e): A aggregated in long horizontal chains and B aggregated
separately. The point-nearest object distribution for Figures (a), (c) and (e) is
shown in Figures (b), (d) and (f) respectively. .............................................. 54
Figure 3-10: Effect of AB clustering on normalized index of dispersion. case 1: A
clustered B clustered separately, case 2: AB clustered , case 3: AB cluster
rearranged. Object A are circles and object B are triangles. ....................... 58
Figure 3-11: Frequency distribution (bins = 40) for froth image shown in Figure 3-5a.. 61
Figure 3-12: Pre-processing (Leo, 2013) of a froth image shot with 40x microscope lens,
(a): color image [produced with permission from Awosemo (2016)], (b):
greyscale image, (c): image (b) after contrast enhancement, (d): image (c)
after homomorphic compression, (e): image (d) after bi-level thresholding,
(f): image (e) with objects touching left and top edge removed, objects
smaller than 2.3µm (20 pixels) removed ..................................................... 63
x